Identifying the Thermal Storage Stability of Polymer-Modified Asphalt with Carbon Nanotubes Based on Its Macroperformance and Micromorphology

.e thermal storage stability of polymer-modified asphalt (PMA) is the key to avoid performance attenuation during storage and transportation in pavement engineering. However, phase separation of PMA continuously occurs after long-term thermal storage due to the overlooked influence of the phase interface. Two kinds of carbon nanotubes (CNTs) and styrene-butadiene-styrene triblock copolymer (SBS) were selected in this paper to address the aforementioned issue..e segregation test was used to simulate the long-term storage process from 0 to 10 days. Macroperformance included the softening point difference (△SP), irrecoverable compliance (Jnr), recovery rate (R%), and complex modulus (G∗) measured by the softening point test, multistress creep recovery (MSCR) test, and small strain oscillatory rheological test. Microcharacteristics were obtained by the SBS characteristic peak index, SBS-rich phase distribution, polymer swelling degree, and particle characteristics of the SBS-rich phase. .ey were measured by Fourier-transformed infrared spectroscopy (FT-IR), fluorescence microscopy (FM), and atomic force microscopy (AFM), respectively. Results showed that the optimum CNTamount necessary to obtain an improved thermal storage stability of PMA was 0.5 wt.%. After 10 days of storage, the largest R% of SBS modified asphalt (SBSMA) decreased to 2.24% and the smallest Jnr increased to 0.0691/kPa, while R% of SBSMA with CNTs was 62.15% and its Jnr was 0.0131/kPa. R% and Jnr of SBSMA with CNTs showed almost no change after 6 days of storage, implying an effective antirutting performance. .e results from the microperformance investigation showed that phase separation of SBSmainly occurred on day 4, while SBS degradation and base asphalt aging led to the worse macroperformance after 10 days of storage. Additional CNTs restrained the SBS-rich phase from floating upward. Meanwhile, a small size of polymer-rich phase and dense network of SBSMA with CNTs were observed in fluorescence microscopy and atomic force microscopy images, thereby exhibiting improved thermal storage stability. Adding CNTs would retard the segregation due to CNT entanglement with SBS.


Introduction
Polymer-modified asphalt (PMA) is regarded as the main pavement material due to its excellent performance compared with the traditional one [1]. However, the inevitable hightemperature transportation and storage remain as challenges for PMA. Each asphalt tank truck can load approximately 40 tons of asphalt. Using up a truck of asphalt in a short time is difficult, especially during bad weather in pavement engineering. erefore, PMA must be stored around 160°C in the asphalt mixing station for a long time [2]. e performance degradation of PMA is inevitable during transportation and storage due to poor thermal storage stability [3]. e most popular polymer modifier is styrene-butadiene-styrene triblock copolymer (SBS), and its marketing share has exceeded 90% worldwide. us, we take the SBS modified asphalt (SBSMA) as an example in this paper. e SBSMA is essentially a thermodynamically incompatible system [4]. At high temperatures, the small particle of SBS is automatically aggregated into the large size and separated from the asphalt phase, thereby causing SBS phase agglomeration and floating on the surface of the asphalt matrix [5]. Finally, the SBSMA performance is deteriorated due to the phase separation of SBS and asphalt [6]. e thermal storage stability of polymer-modified asphalt captures the performance change during high-temperature conditions. erefore, this performance has been getting a lot of attention.
Many methods are used to improve the thermal storage stability of SBSMA. Mouillet et al. reported that gravity is the direct cause of phase separation [7].
us, changing the density difference between SBS and asphalt would improve the thermal storage stability. Wang et al. indicated that adding aromatic oil would help obtain effective compatibility of asphalt and SBS because aromatic compounds in asphalt had a strong interaction with polybutadiene in SBS [8]. However, Sun and Lu studied and interpreted the different hot storage stabilities of three SBS modified asphalts with the interface layer principle; the results show that physical modification methods cannot work because SBS interphase still relied on van der Waals forces to form the physical network [9]. Lesueur reported that the physical network of SBS would be destroyed by heating or adding external force, thereby leading to the agglomeration and separation of the SBS-rich phase from the asphalt phase [10].
us, some chemical methods were proposed. e most widely used method is the sulfur crosslinking agent. However, Zhu et al. revealed that sulfur cross-linked SBS modified asphalt was sensitive to oxygen and easily leads to aging and bond breaking [11]. e thermal separation of SBSMA remains unsolved if only the sulfur crosslinking agent is used. Wang et al. demonstrated that the weak interphase between SBS and asphalt resulted in poor thermal storage stability based on polyphase material mechanics and with the help of the Palierne model [12]. Zhu et al. established a PMB phase field model considering viscoelasticity and calibrated it with the experimental observations of three different PMBs; they reported that the separation of SBSMA was subjected to interphase characteristics [13]. Carbon nanotubes (CNTs) were used to improve the interphase strength of composite and increase asphalt performance [14,15]. However, the effect of different kinds of CNTs on the long-term thermal storage stability of SBSMA was unclear.
is paper aims to investigate the improvement of thermal storage stability for SBSMA with or without CNTs. Multiwalled CNTs were selected, and two ways of adding CNTs in the asphalt were proposed. e segregation test was used to simulate the thermal storage procedure. Macroperformance was obtained from the softening point test, multistress creep recovery test, and small strain oscillatory rheological test. Microcharacteristic indexes were captured by FT-IR, FM, and AFM. Here, microcharacteristics were used to elucidate the thermal storage performance changes with time, including SBS characteristics peak index, SBS-rich phase distribution and swelling degree, and SBS-rich phase particle characteristics. e work would provide some new ideas to identify the polymer-rich phase segregation of polymer-modified asphalt completely.

Materials.
e raw material included straight asphalt, SBS polymer, CNTs, and stabilizer. Straight asphalt with penetration of 70 dmm was used as base asphalt (marked as A70#); the technical properties are shown in Table 1. Linear type SBS with molecular weight of 110000 g/mol was used in this research, and the block ratio of styrene to butadiene is 3 : 7. CNTs were surface-modified with hydroxyl. Stabilizer was a type of commercial sulfur crosslinking agent. It is composed of 80% sulfur powder and 20% filler, in which the filler is nano calcium carbonate.
Two types of CNTs were selected, marked as CNTs-I and CNTs-II, respectively, as seen in Figure 1, and the technical properties are shown in Table 2. CNTs-I are the original multiwalled carbon nanotubes with hydroxyl, the length of CNTs is 10-30 μm, and the relative content of functional groups is 2.1 mmol/g. CNTs-II refer to the SBS polymer alloy with original multiwalled carbon nanotubes containing hydroxyl.

Preparation of Polymer-Modified Asphalt Samples.
Polymer-modified asphalt samples were prepared using a high-shear mixer (WeiYu Machine Co., Ltd, Shanghai, China). Base asphalt was heated at 160°C for improved flow. SBS and CNTs were added to the mixture of asphalt at 180°C for 30 min at a fixed rotation speed for 3000-3500 r/min. ereafter, asphalt stabilizer (0.25 wt.% by asphalt weight) was added into the mixtures at 175°C for 5 min at a fixed rotation speed of 2500 r/min. Finally, the mixture temperature was decreased to allow the polymer swelling and prevent asphalt aging.
Moreover, CNTs-I and CNTs-II showed similar adding methods into the asphalt. However, CNTs-I were blended with SBS particle in advance, while CNTs-II were directly added into the heated base asphalt mixed with SBS particle. e obtained asphalt was marked as SBSMA, SBS + CNTs-I, and SBS + CNTs-II. Herein, SBSMA referred to polymermodified asphalt only using SBS particle, SBS + CNTs-I contained SBS particle and CNTs-I, and SBS + CNTs-II contained SBS and CNTs-II. All polymer-modified asphalts satisfied the relevant requirements of I-C grade modified asphalt according to the "Technical Specification for Highway Asphalt Pavement Construction" (F40-2004) in China. e samples used in the test are shown in Table 3.

Experimental Methods.
e storage samples used in the test were obtained from the segregation test in accordance with ASTM D 5975. e process of segregation test is shown in Figure 2. However, the storage time was not only 48 h. In the CNTs content determination part, the segregation time was 48 h. e SBS + CNTs-I samples were prepared for testing with 0 wt.%, 0.2 wt.%, 0.5 wt.%, and 1 wt.% of CNTs-I combined with a solid 4.0 wt.% SBS by weight of asphalt. Meanwhile, the SBS + CNTs-II samples were prepared for testing with 0 wt.%, 0.5 wt.%, 1.5 wt.%, and 2.5 wt.% of 2 Advances in Materials Science and Engineering CNTs-II combined with a total of 4.0 wt.% SBS by weight of asphalt. e effective CNTs in CNTs-II were the same as those in CNTs-I. us, the amount of CNTs-II was similar to CNTs-I in subsequent sections. e experimental design flowchart is shown in Figure 3. e macroperformance of SBSMA was obtained by the softening point test, multistress creep recovery (MSCR) test, and small strain oscillatory rheological test. e softening point difference (△SP) was obtained by measuring the softening point at the top and bottom of the sample by the global method in accordance with ASTM D5976. MSCR was used to capture irrecoverable compliance (J nr ), recovery rate(R%), and accumulative strain in accordance with ASTM D7405-10a. e small strain oscillatory rheological test was used to get the complex modulus (G * ) on the dynamic shear rheometer (DSR) with a temperature at 70°C. MSCR and small strain oscillatory rheological tests are carried out on the dynamic shear rheometer (DSR). e test principle is that the rotation axis applies periodic sinusoidal shear stress to the asphalt sample, from point A to point B, then back to point A, then to point C, and finally back to point A, which is completed in one cycle, as shown in Figure 4. e microscopy tests were conducted using FT-IR, FM, and AFM. FT-IR was used to capture the SBS characteristic peaks changes, which was performed on TENSOR II in the storage process of SBSMA, the number of tracing is 64 cm −1 , and the resolution is 4 cm −1 . FM was used to observe the SBS-rich phase distribution, and AFM was used to identify the particle characteristics of SBS-rich phase. FM was conducted on Leica DM 2500, and the magnification is 100x. AFM was performed by the Bruker Dimension Fast Scan, and the sample test range is 50 μm × 50 μm, 10 μm × 10 μm, and 3 μm × 3 μm. e scanning frequency is 1 Hz, and the scanning resolution is 256 frames × 256 frames. Digital image technology was used to find the SBS swelling degree and particle characteristic indexes. e asphalt films for FM and AFM were obtained by employing a heat-casted method. A hot asphalt binder was cast into glass slides to achieve effective flow with a heating temperature of approximately 160°C. Bitumen-covered sample holders were left overnight at room temperature before testing.
In addition, the degree of separation (△ parameters) was used to capture the performance differences between the top

Determination of CNT Amount in SBSMA.
CNT amount plays a crucial role on the thermal storage property of SBSMA. e standard segregation test was used to obtain the 48 h storage asphalt samples with different CNT amounts. e conventional index (△SP) and rheological indexes (R%, J nr , and G * ) were used to identify the thermal storage stability changes of SBSMA with CNT amount during storage procedure.
△SP was a common index used to evaluate the segregation degree and the performance deterioration of PMA. A large △SP demonstrated poor thermal storage stability. In Figure 5, the samples with CNTs all satisfied the requirement of specification in China. For the sample of SBS + CNTs-I, △SP decreased at first and then increased, yet the smallest △SP was observed at 0.5 wt.%. is phenomenon was ascribed to the addition of appropriate amount of CNTs can improve the stability of SBSMA, but the accumulation of a large number of CNTs in asphalt will lose the reinforcement capability of asphalt. For the sample of SBS + CNTs-II, △SP decreased with CNTs-II amount. No peak was observed in the curves of △SP in Figure 5(b). However, another study pointed out that this conventional index cannot express segregation accurately [4].
us, additional investigations are necessary to find the optimum dosage of CNTs-II in SBSMA.
e MSCR test was selected to identify R and J nr changes with increasing CNT amount, and △R and △J nr were, respectively, calculated from equation (1), as indicated in the results in Figure 6. For the sample of SBS + CNTs-I, Figures 6(a) and 6(b) revealed that R% at the top initially increased and then later decreased, while the value of R% at the bottom initially decreased and then increased. J nr of SBS + CNTs-I presented a similar rule to R%. For the sample of SBS + CNTs-II, the rule of R% was unclear, but the largest R % and the smallest J nr were obtained at the additional 0.5 wt.% CNTs-II. A large R and small J nr revealed a remarkable  ) SBS-rich phase distribution SBS swelling degree SBS-rich phase particle characteristics  Advances in Materials Science and Engineering antirutting performance of PMA [10]. us, additional 0.5 wt.% CNTs would help obtain effective high-temperature performance. Meanwhile, the results in Figures 6(c) and 6(d) showed that △R decreased before CNTs were below 0.5 wt.% and increased when CNTs surpassed 0.5 wt.%. For the addition of 0.2 wt.% CNTs, △R of SBSMA with CNTs-I was larger than that with CNTs-II, while the rule of △J nr was completely opposite. e value of △J nr for SBS + CNTs-I initially decreased and then increased. By contrast, the value for additional CNTs-II initially increased and then decreased, thereby finally raising the CNT amount. is phenomenon might be subject to the discrepancy in the interaction between CNTs and SBS among the asphalt matrix. e small △R and △J nr illustrated a minimal difference between the top and bottom of SBSMA during the storage process.
us, the optimum amount of CNTs was 0.5 wt.% from the hightemperature performance of SBSMA ( Figure 2). Figure 7 provides the complex modulus (G * ) variation of SBSMA with different types of CNTs. For SBS + CNTs-I, G * significantly decreased with CNTs-I amount at the top of asphalt samples in Figure 7(a). By contrast, G * initially decreased but later increased at the bottom with increasing CNTs-I amount. For SBS + CNTs-II, G * at the top slightly decreased, while the value of G * at the bottom almost remained constant with increasing CNTs-II amount.
Compared with CNTs-I and CNTs-II, the samples with 0.2 wt.% CNTs-II at the top had the smallest modulus, while the one with 1.0 wt.% CNTs-I at the bottom had the largest modulus. A large modulus would provide improved high-temperature strength [6]. us, the modulus changes in asphalt indicated that the top and bottom of SBSMA had different high-temperature deformation resistance. Figure 7(b) revealed that △G * initially decreased and then increased with rising CNTs-I amount, while the one with CNTs-II was different. However, the smallest △G * was found in the amount at 0.5 wt.% for both types of CNTs.
Overall, the data of △SP, △R, △Jnr, and △G * for SBSMA with different types of CNTs had different rules for the thermal storage process. However, the smallest change between the top and bottom for those indexes was an additional 0.5 wt.% CNTs. us, the optimum percentage for adding CNTs into SBSMA was 0.5 wt.% for further study.  decreased. Although SBSMA satisfied the requirement of specification on the 2nd day, the value of △SP on the 4th day was beyond this requirement. is trend was ascribed to SBS-rich phase distribution characteristics, base asphalt aging, and SBS degradation in high-temperature conditions [16]. All these factors led to poor network strength of SBS. However, modifying this essential characteristic of the thermodynamic incompatible system is difficult. Comparatively, adding CNTs would improve the degree of deterioration. e SBS + CNTs-I showed more gentle differences of △SP than that of neat SBS. Meanwhile, the maximum value of △SP for SBS + CNTs-I decreased, thereby demonstrating that adding CNTs-I could decrease the segregation degree of SBSMA. e SBS + CNTs-II had no significant advantage in decreasing △SP of SBSMA possibly due to the large cohesive interaction between CNTs and SBS in CNTs-II. However, additional work proved that △SP was an ineffective index to evaluate the segregation degree [4]. us, the creep characteristics and rheological performance were used for further research.

Creep
Characteristics from MSCR Test. R, Jnr, and accumulated strain of three modified asphalt were selected to capture the creep characteristic differences of SBSMA before and after adding two types of CNTs at long-term thermal storage period, as shown in Figures 9 and 10. Figure 9 shows that the values of R% for SBSMA either in the top or bottom decreased with increasing storage time. e R% value of SBSMA was close to zero after 10 days of thermal storage. e values of J nr for SBSMA increased with storage time. e samples with additional CNTs showed a larger R% and smaller J nr than that without CNTs. A large R% represented that asphalt performed small permanent deformation  Advances in Materials Science and Engineering caused by the viscous flow and showed low risks of rutting in asphalt pavement [17]. A small J nr means that the asphalt had minimal irrecoverable residual strain and produced small rutting depth in asphalt pavement [17]. us, the results in Figure 9 illustrated that additional CNTs could improve the antirutting performance of SBSMA regardless of the CNTs type. Meanwhile, R% and J nr of the samples with additional CNTs were almost maintained when the time was less than 8 days. is phenomenon demonstrated that additional CNTs could prolong the thermal storage stability to 8 days. Comparing the CNTs-I and CNTs-II, the storage time was less than or equal to 6 days, and the improvement effect of thermal storage stability was similar. Meanwhile, the CNTs-II had better antirutting performances than those with CNTs-I when storage time was higher than 6 days. Figure 10 shows the accumulative strain of three modified asphalts after 10 days of thermal storage. Either at the top or bottom, the SBSMA showed the largest accumulative strain, the SBS + CNTs-II had the smallest accumulative strain, and the SBS + CNTs-I was between the two. is finding might be ascribed to the distribution of the SBS-rich phase and SBS amount changes in the top and bottom of the asphalt matrix. Combined with the data in Figure 9, neat SBSMA lost its elastic recovery capability after 10 days of storage. Adding CNTs-I into SBSMA, the accumulative strain was reduced, the elastic recovery capability was improved, and the elastic recovery capability of the top and bottom parts had minimal difference. Adding CNTs-II into SBSMA, the accumulative strain of the top was reduced to approximately 1/2 of that of SBSMA, and the value of the bottom reached 1/5 of that of SBSMA. us, the results in Figures 9 and 10 revealed that additional CNTs would endow the SBSMA with outstanding elastic recovery capability even after 10 days of thermal storage, and the CNTs-II showed better improvement than that with additional CNTs-I.

Rheological Characteristics from DSR.
Complex modulus (G * ) is an important rheological index, which is closely related to high-temperature performance calculated from the stress-strain ratio of viscoelastic materials subjected to sinusoidal loading. us, G * was selected to investigate the rheological characteristics of three modified asphalts, as shown in Figure 11. Figure 11 provides the G * changes of three modified asphalts with increasing storage time. For SBSMA, G * at the top and bottom substantially decreased with increasing   Advances in Materials Science and Engineering storage time. △G * on the 2nd day of SBSMA was also larger than that with CNTs. e attenuation rate for the sample of SBSMA containing CNTs was significantly slower than that of SBSMA without CNTs. Compared with the SBS + CNTs-I and SBS + CNTs-II, additional CNTs-I into SBSMA could remarkably minimize the decreasing G * at the top if storage time surpasses 6 days, while G * at the bottom was increasing if the storage time is beyond 2 days. Additional CNTs-II into SBSMA showed similar rules, yet the change degree was not the same in Figure 11(a). is phenomenon demonstrated that the reinforcement of CNTs improved the modulus changes with the degradation of the SBS network in asphalt. Moreover, the modulus was slightly decreased when the storage time surpassed 8 days. is phenomenon illustrates that additional CNTs could slow down the decay of G * during long-term thermal storage and provide improved high-temperature performance. Meanwhile, the CNTs-II showed better improvement than CNTs-I in SBSMA.

SBS Characteristic Peak Index Changes from FT-IR.
Characteristic peaks in the FT-IR spectrum provided the functional group composition in the substance. Figure 12 summarizes the FT-IR spectra of the three modified asphalt samples with different storage times. e peaks are divided into two sections in the FT-IR spectra: functional group area (1000-4,000 cm −1 ) and   Advances in Materials Science and Engineering fingerprint area (400-1000 cm −1 ). In the functional group area, the observed wavenumbers at 2850, 2920, and 1450-1475 cm −1 demonstrated the existence of an aliphatic long chain in saturated hydrocarbon [18]. In the fingerprint area, SBS showed characteristic peaks at wavenumbers of 699, 760, 910, and 966 cm −1 , which belonged to benzene or aromatics [19]. e peaks at 699 and 966 cm −1 belonged to the polystyrene phase of SBS [20,21]. e 966 cm −1 , which was the butadiene stretching vibration of SBS, has a stronger peak intensity of functional groups in SBS than 699 cm −1 . e base asphalt characteristic peaks included 1377, 1450-1475, 2850, and 2920 cm −1 . 1377 cm −1 represented the base asphalt [22]. e SBS characteristic peak index I SBS was used in this study to capture the SBS amount changes in asphalt matrix, as shown in equation (2), where S 966 and S 1377 were the peak areas of 966 and 1377 cm −1 , respectively. e smaller the difference between the top and bottom I SBS value, the better the thermal stability of modified asphalt.   is finding showed similar rules for SBS + CNTs-II, thus revealing different I SBS . Figure 13 provides the I SBS and separation degree changes based on I SBS in storage for 10 days. I SBS of SBSMA at the top increased in the last 4 days and then decreased  until 10 days in Figure 13(a). A large I SBS illustrated a large SBS amount in the asphalt. us, the rules demonstrated that SBS particles floated up with the increase in storage time. is phenomenon resulted in a large SBS content in the top but a small SBS amount in the bottom due to a smaller density for SBS than asphalt. e results also demonstrated that the phase separation mainly occurred within 4 days and remained stable in the last 6 days. e indexes of I SBS for the samples with CNTs were slightly changed during 10 days of storage. Comparing the two types of CNTs, I SBS of the samples with CNTs-I at the top revealed almost no changes with storage time, while the one with CNTs-II was slightly decreased. e index of ISBS for samples with CNTs-I or CNTs-II increased with storage time at the bottom. is phenomenon revealed that the increase in SBS amount at the top of all samples was reduced by adding CNTs, while the slightly increased values at the bottom of all samples proved that the network between CNTs and SBS was insufficiently strong. Moreover, the network structure formed by the mixture of CNTs and SBS would lead to an increase in polymer density and facilitate the easy settlement of SBS to the bottom [23].
However, △ I SBS indicated a different result. In Figure 13(b), the △ISBS for the samples with CNTs increased with storage time, and the one with CNTs-II showed large values at every point. e particle size of CNTs-II was significantly larger than that of CNTs-I. us, an evident increase in polymer density of CNTs-II and SBS leads to the easier settlement of SBS particle to the bottom than that of CNTs-I. △ I SBS of SBSMA was completely different, which increased in the last four days and then decreased until 10 days. Although △ I SBS of SBSMA decreased on the 10th day, the performance of SBSMA had been seriously damaged. A crusting phenomenon was observed on the 10th day of the SBSMA, demonstrating that SBS polymer was completely destroyed after 10 days of high-temperature storage. us, the difference softening point only between the top and bottom was not fit to capture the phase separation of polymer-modified asphalt.

Distribution and Swelling Degree of SBS-Rich
Phase from FM. FM was used to capture the SBS-rich phase distribution characteristics for SBSMA with CNTs as listed in Figure 14. e FM images revealed that the degree of SBS swelling was calculated on the basis of equation (3) using digital image technology, which was used to estimate the compatibility of polymer and asphalt indirectly, as proposed by Sengoz [24]. In the equation, A PRP is the area of SBS-rich phases, and A total is the total area of SBSMA in FM images. e original FM images were processed with MATLAB to convert them into binary images, which only contained two-pixel values (namely, 0 and 1), and then to binary black (asphaltene-rich phases) or white (SBS-rich phases). A PRP and A total are finally obtained [25]. e degree of SBS swelling is listed in Figure 15.

Degree of SBS swelling �
A PRP A Total × 100%.
(3) Figure 14 shows that the SBS-rich phases are bright yellow, while asphaltene-rich phases and CNTs are almost black due to the absence of fluorescence. For SBSMA, the number of particles and the particle size in the top part all increased, while their behaviours were completely opposite in the bottom as shown in Figure 14(a). ese results were verified by Figure 13(a). For the samples with CNTs, the SBS-rich phase showed a smaller size and denser network than that of SBSMA. For SBS + CNTs-I, the particle size of the SBS-rich phase was slightly larger on the 10th day than that of the 2nd day either on the top or bottom. Meanwhile, the bottom of SBS + CNTs-I on the 10th day showed a considerable amount of SBS particles than that at the top.
e rule was the same as the results in Figures 10 and 11 from FT-IR. For SBS + CNTs-II, the SBS network was substantially denser than that with CNTs-I. Even on the 10th day, the large SBS particles were still less than those in the SBS + CNTs-I. Figure 15 shows the degree of SBS swelling for three modified asphalts with increasing storage time. For SBS modified asphalt, the degree of SBS swelling at the top was increased, yet these values sharply decreased with increasing storage time. is result demonstrated that with the increase of storage time, SBS floats on the asphalt surface, and the amount of SBS dissolved in the top significantly increased while that in the bottom decreased, finally leading to phase separation or segregation of SBSMA. After 4 days of storage, the phase separation gradually stabilized, but a large amount of SBS floats up, resulting in some SBS slowly precipitating from the asphalt and crust on the surface with the passage of time. us, the content of SBS in the bottom slightly decreased in the last 6 days, as the results from FT-IR in Figure 13. A downward trend of the bottom in Figure 15(a) was due to the severe phase separation initially and then gradual gentleness with the increase in time. e samples with CNTs showed a substantially denser SBS network than that of neat SBSMA. For the SBS + CNTs-I, an upward trend either in the top or bottom was observed with increasing storage time in Figure 15(b), thereby providing improved compatibility between SBS and asphalt to avoid separation. Wang attributed this increase to the channelling effect of CNTs to transport small molecules in asphalt and improve the compatibility between SBS and asphalt [26]. e SBS + CNTs-II showed an upward trend in the bottom in Figure 15(c), but a downward trend was observed in the top with increasing storage time. Despite the poor compatibility of the top of SBS + CNTs-II, the data were still larger than that of SBSMA.
us, the results showed that additional CNTs could improve the thermal storage stability of SBSMA. However, the results from FM images unclearly identified the difference between CNTs-I and CNTs-II. erefore, the AFM was selected to clarify this issue.

Particle Characteristics of SBS-Rich Phase from AFM.
e particle characteristics of the SBS-rich phase for SBSMA with CNTs during long-term storage were captured from AFM, as presented in Figures 14 and 15. In AFM images, the dispersed phase showed some tiny white dots mainly comprising the SBS-rich phase, while the continuous phase was the asphalt matrix. e bee structures of the base asphalt were ascribed to wax, asphaltene, and trace metals in asphalt [27,28]. When adding SBS into asphalt, the bee structures were enclosed by swelled SBS, finally becoming blurred in AFM images [29]. us, the SBS-rich phase was used to identify the SBS network changes with increasing storage time.
For SBS + CNTs-I in Figure 16, determining the SBSrich phases and asphalt matrix at the zero day is difficult, but the famous bee structure was still observed in the tiny white dots. After storage of 4 days, the size of the bee structure was slightly larger in the bottom than that in the top of SBS + CNTs-I. e images on the 8th day of SBS + CNTs-I revealed that the bee structure at the bottom was significantly larger than that at the top, thereby demonstrating the phase segregation occurrence. e images on the 10th day of SBS + CNTs-I indicated that the clear bee structure was observed at the top and bottom, thereby implying the degradation of SBS polymers. e bee structure size at the top was slightly larger than that at the bottom. For the sample of SBS + CNTs-II in Figure 17, the SBS-rich phase of SBS + CNTs-II at zero days was denser than that of SBS + CNTs-I. After storage of 10 days, the SBS network of SBS + CNTs-II at the top and bottom revealed insignificant deterioration.
us, the surface roughness based on equations (4)-(6) from the nano scope analysis based on digital image technology was used for the quantitative description of the particle characteristics discrepancy. In the equation, R q is the root mean square of roughness, R a is the roughness average, Z(x) is the function that describes the surface profile analysed considering height (Z) and position (x) of the sample over the evaluation length "L," and R max is the maximum roughness.   14 Advances in Materials Science and Engineering Table 4 summarizes the surface roughness of asphalt samples with different storage times. For SBS + CNTs-I, R q and R a at the top with 2 days of storage were gradually increased due to base asphalt thermal aging and initial aggregation of small SBS particles to the large ones. In 2-4 days, the values of R q and R a started to decrease due to the absorption of some CNTs-I into SBS, easily sinking the large density of the dispersed phase to the bottom. In 4-8 days, the aggregation degree continued to worsen for SBS + CNTs-I. In 8-10 days, the aging for the base asphalt led to a crusting phenomenon and provided the largest R for the SBS + CNTs-I. For SBS + CNTs-II, R q and R a at the top slowly increased and slightly decreased. However, the increase in Rmax at the top was ascribed to the aggregation of the SBS-rich phase within 10 days. At the bottom of SBS + CNTs-II, although fluctuations in the values for R q and R a were observed, these fluctuations were small. erefore, the overall situation was relatively stable. e overall trend of R max slowed down, thereby illustrating that some SBS particles were still floating to the top part after 10 days of storage. Zhai et al. reported the polymer would entangle with carbon nanomaterials [30].
us, the phenomenon here would be ascribed to CNTs entangling with SBS. However, CNTs-II showed better inhibition effect on the aggregation of SBS particles than CNTs-I.

Conclusions
e elusive thermal storage stability of SBSMA has attracted wide attention for over half a century of intensive research. However, phase segregation constantly occurs after long-term thermal storage for SBSMA.
us, this study aimed to investigate this issue using CNTs due to their excellent interface enhancement. e role of CNT amount on the performance of SBSMA was investigated.
e results revealed that SBSMA with different types of CNTs had different rules to the thermal storage stability. e optimum amount of CNTs for improving storage stability of SBSMA was 0.5 wt.%. is condition resulted in the smallest changes between the top and bottom for softening point difference, irrecoverable compliance, recovery rate, and complex modulus.  e macroperformance of SBSMA with CNTs within 10 days of storage was examined. e results revealed that for the difference of softening point, CNTs-II had no significant advantage in decreasing △SP of SBSMA but CNTs-I had. For rheological characteristics, after 10 days of storage, the largest R% of SBS modified asphalt (SBSMA) decreased to 2.24% and the smallest J nr increased to 0.069 1/kPa, while R% of SBSMA with CNTs was 62.15% and its J nr was 0.013 1/kPa. R% and J nr of SBSMA with CNTs showed almost no change after 6 days of storage. erefore, adding CNTs endowed the SBSMA with outstanding elastic recovery capability even after 10 days of storage. Moreover, adding CNTs slowed down the decay of complex modulus, and the CNTs-II showed better improvement than CNTs-I in SBSMA. e microperformance of SBSMA with CNTs within 10 days storage was investigated by FT-IR, FM, and AFM. e results showed that SBS floated up in SBSMA but sank after adding CNTs. It was due to the entanglement between CNTs and SBS, which increases the density of SBS phase. Additional CNTs altered the distribution of SBS-rich phase and enhanced the swelling degree during the storage process. A small and dense network of SBS with CNTs was observed in FM and AFM images, thereby exhibiting improved thermal storage stability. Moreover, CNTs II worked better than CNTs I for improving storage stability of SBSMA; therefore, adding CNTs directly into SBSMA demonstrated less modification than the one premixed with SBS in advance.
is study provided a new method to improve the thermal storage stability of polymer-modified asphalt.

Data Availability
e data used to support the findings of this study are included within the article.

Disclosure
is study was completed at the School of Transportation Engineering in Shandong Jianzhu University and at the School of Transportation Science and Engineering in Harbin Institute of Technology.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.